Public Sector Employment and the Skill Premium: Sweden versus the United States 1970–2012

• DOI: http://doi.org/10.1111/sjoe.12284
• Published date: 01 January 2019
• Date: 01 January 2019

Scand. J. of Economics 121(1), 3–31, 2019

DOI: 10.1111/sjoe.12284

Public Sector Employment and the Skill

Premium: Sweden versus the United States

1970–2012*

David Domeij

Stockholm School of Economics, SE-113 83 Stockholm, Sweden

david.domeij@hhs.se

Lars Ljungqvist

Stockholm School of Economics, SE-113 83 Stockholm, Sweden

lars.ljungqvist@hhs.se

Abstract

Swedish census data and tax records revealan astonishing decline in the agg regateskill premium

of 30 percent between 1970 and 1990, with only a modest recoveryin the next couple of decades.

In contrast, the US skill premium rose by around 24 percent over those four decades.A theor y

that equalizes wages with marginal products can rationalize these disparate outcomes when we

replace commonly used measures of total labor supplies by private sector employment. The

dramatic decline in the skill premium in Sweden is the result of an expanding public sector that

has disproportionately hired unskilled labor.

Keywords: Private versus total employment; skill-biased technological change; skilled and

unskilled labor; wage compression

JEL classification:E24; J31

I. Introduction

It is well known that Sweden experienced a dramatic wage compression

after the 1960s, with accompanying lower returns to skills. In this paper,

we seek to explain those lower returns by, first, computing as accurately

as possible a measure of the aggregate Swedish skill premium based on

population data and, second, using a structural model to rationalize the

skill premia of Sweden and the United States in terms of observed labor

quantities between 1970 and 2012.

Using census data and tax records that cover the entire working-

age population, we document a spectacular collapse of the Swedish skill

*We are grateful to three anonymous referees and numerous seminar participants for helpful

comments. Financial support from the Jan Wallanderand Tom Hedelius Foundation is gratefully

acknowledged.

C

The editors of The Scandinavian Journal of Economics 2017.

4Public sector employment and the skill premium

premium. The Swedish skill premium fell by 30 percent between 1970 and

1990, while the US skill premium, after an initial decline in the 1970s, rose

by around 8 percent. Since then, the Swedish skill premium has slightly

rebounded while the US skill premium has continued to increase by another

16 percentage points in 2012. These vastly different outcomes are not a

result of any significant difference in initial wage differentials; for example,

Edin and Topel (1997, p. 173) report that the typical college graduate in

1968 earned 50 percent more than an otherwise comparable high school

graduate in both the United States and Sweden.1

The dramatic compression of the Swedish wage distribution has led

many observers to question whether Swedish workers are still paid their

marginal products. For example, Edin and Topel (1997) theorize that

Swedish wages have become unhinged from marginal products because

of collusion between the trade union and the employers’ association.2We

will instead explore whether marginal products of aggregates of skilled and

unskilled labor can explain the evolution of the Swedish skill premium.

We use a production function with a capital–skill complementarity as

formulated by Krusell, Ohanian, R´ıos-Rull, and Violante (2000), hereafter

denoted KORV. KORV showed that the historical increase in the relative

supply of skilled labor in the United States was more than offset by

skill-biased technological change, as measured by falling quality-adjusted

equipment prices, which enabled them to successfully account for the rising

US skill premium between 1963 and 1992.

As in earlier cross-country studies of skill-biased change (e.g.,

Gottschalk and Joyce, 1998; Acemoglu, 2003), we find that total

employment of skilled and unskilled labor cannot explain the Swedish

skill premium. However, we also find that private employment can. Our

analysis suggests that the dramatic decline of the skill premium in Sweden

is the result of an expanding public sector that has disproportionately

hired unskilled labor. The excessive hiring of unskilled labor in the public

sector propagated a relative scarcity of unskilled labor in the production

function of the private sector, which pushed the skill premium down. This

1For contemporary evidence of such high rates of return to university education in Sweden in

the late 1960s, see the government commissioned report by St˚ahl (1974). However,rather than

specific wage differentials, our study of the aggregate skill premium is cast in terms of a skill

premium index by which the wages of all unskilled and all skilled workers, respectively, are

weighed together so as to deduce how the wage perunit of skilled labor has changed over time

relative to the wage per unit of unskilled labor (as in the US studies by Katz and Murphy, 1992;

Krusell et al., 2000).

2In the model of Edin and Topel(1997, pp. 192–197), the egalitarian objectives of the trade union

coalesce with the interests of the employers’association because the bargained “agreements did

not just raise the compensation of low-wage workers; they also reduced the absolute wage of

skilledworkers...delivering‘cheap’skilledlabor to large employers.”

C

The editors of The Scandinavian Journal of Economics 2017.